Endometriosis is a female reproductive disorder characterized by the presence of endometrial tissue outside of the normal uterine location. Most frequently the endometriosis tissue is present in the peritoneal cavity, attaching to various tissues and organs in this location. Endometriosis is a benign disease affecting approximately 5 million women in the United States annually with a prevalence of 10 to 15 percent in women of childbearing age. The incidence increases to 60 to 80 percent of women who are infertile or present with pelvic pain (D. Gosselin et al. [1999] Curr. Opin. Onco. Endo. & Metabol. Invest. Drugs 1:31). The conditions that predispose an individual to endometriosis are still unknown. Several authoritative reports suggest that retrograde menstruation may be a key-contributing factor, but this process is thought to be common in most women. This theory has also been questioned recently (D. B. Redwine [2002] Fert. Steril. 78:686) due primarily to the substantial differences that occur between normal or eutopic endometrium and the ectopic tissue found in diseased patients. Consequently, other genetic as well as immunological factors are thought to contribute key roles to the development of the disease in susceptible women. For example, endometriosis is thought to be much more frequent in first degree relatives of affected women when compared to the rest of the population (Coxhead and Thomas [1993] J. Obstet. Gynacol. 13:42). In addition to the frequency, the disease has also been reported to be more severe in women with a first-degree relative with endometriosis (Thomas and Campbell [2000] Gynacol. Obstet. Invest. 50:2). The precise gene(s) involved in the disorder are unknown but the pattern is strongly suspected to be maternal in nature.
Although not life threatening, endometriosis results in substantial abdominal discomfort, and may cause infertility. In fact, such symptoms can be indicative of other feminine health disorders and this makes the diagnosis of endometriosis clinically challenging. This was emphasized in a recent study upon the effects of delayed diagnosis of endometriosis (G. K. Husby et al [2003] Acta. Obstet. Gynecol. Scand. 82:649). These investigators reported delays from 3 to 11 years between the onset of pain and the final diagnosis of endometriosis. In this study women reporting both infertility and pain did not have a significantly shorter delay in diagnosis. Obviously such delays coupled with the symptoms reported lead to the expenditure of considerable economic and psychological resources.
Currently surgical laparoscopy is considered to be the gold standard for diagnosis of endometriosis. During laparoscopy the disease is visually staged using a point system from stage I (minimal disease, 1 to 5 points) to IV (severe disease, >40 points). The points are assigned according to several parameters such as location, size, and depth (superficial versus deep) of the lesions (T. P. Canavan and L. Radosh [2000] Postgrad. Med. 107:213). Some opinions reveal potential hazards with the procedure, and frequently laparoscopy does not result in a definitive diagnosis of the disease (S. Pillai et al. [1996] Am. J. Reprod. Immunol. 35:483). For example, while laparoscopy is not classed as major surgery it still has several features (invasive, expensive, requires anesthesia, and full operating facilities) which together make the process an unfortunate choice for diagnosis at least. In fact, while endometriosis is not a fatal disorder, laparoscopy itself can be life threatening. The trans-abdominal approach has been reported to be responsible for 50% of the complications from this procedure, and injury to major blood vessels can result in mortality of 15% (I. A. Brosens and J. J. Brosens [2000] Eur. J. Obstet. Gynecol. Reprod. Biol. 88:117). Other complaints are that the visual staging of the disease does not correlate with the degree of infertility or the severity or number of symptoms (T. P. Canavan and L. Radosh [2000] Postgrad. Med. 107:213). It has been reported that the place of laparoscopy in the diagnosis of endometriosis should be reassessed (I. A. Brosens and J. J. Brosens [2000] Eur. J. Obstet. Gynecol. Reprod. Biol. 88:117). Rational for this lies in part due to the suggestion by some (P. R. Koninckx [1994] Hum. Reprod. 9:2202) that mild endometriosis is not a disease at all and that all women have endometriosis. Moreover as noted above there are functional aspects (e.g., infertility, abdominal pain, etc.) to stages other than mild disease and these are more commonly being applied to diagnosis. Consequently it has been proposed that the traditional ‘gold standard’ be replaced with a combination of transvaginal hydrolaparoscopy (THL, a somewhat milder procedure) and magnetic resonance (MR) imaging until suitable biochemical markers have been identified (I. A. Brosens and J. J. Brosens [2000] Eur. J. Obstet. Gynecol. Reprod. Biol. 88:117).
The frequency of endometriosis and the difficulty of diagnosing the disorder together represent ample rationale for experiments designed to identify these serum based biochemical markers. Other discovery phase programs have implied that of potential markers of endometrial disease may exist. For example, levels of the epithelial ovarian-derived antigen CA-125 have been reported to be elevated in serum, peritoneal fluid, and menstrual fluid of endometriosis patients (B. Mol et al. [1998] Fertil. Steril. 70:1101). The marker exhibited good specificity, but the sensitivity is poor with high levels present in patients afflicted with PID, ovarian cancer, or cervical carcinoma. Despite the limitations, the marker may be of use for patients who are likely to have the disease for faster orientation toward laparoscopy, since CA-125 levels do correlate somewhat with the degree of disease and response to treatment (T. P. Canavan and L. Radosh [2000] Postgrad. Med. 107:213).
Also, Sharpe-Timms et al. (Biol. Reprod. [1998] 58:988) have reported that endometriosis lesions secrete a haptoglobin-like protein in a rodent model system. The haptoglobulin was specifically synthesized by endometriosis tissue and was not found in uterine tissue using a sensitive reverse transcriptase PCR technique. This antigen is also interesting in that it has been reported to modulate immune cell functions and could contribute to the pathophysiology of endometriosis
Along slightly different lines, D. Gosselin et al. (Curr. Opin. Oncol. Endo. Metabol. Inv. Drugs [1999] 1:31) reported a diagnostic algorithm employing several different combinations of leukocyte markers present upon subsets of T and B cells, macrophages, and NK cells in peripheral blood and endometrium of patients with endometriosis. This formed the foundation for the development of a diagnostic test (Metrio Test) by PROCREA BioSciences, Inc. which is approved by Health Canada. The Metrio Test is based on the assessment of eight proprietary leukocyte subsets by flow cytometry analysis combined with a blood biochemical marker evaluated by ELISA (J. Brosens et al. Obstet. Gynecol. Clin. North Am. [2003] 30:95-114). This test reportedly has a specificity rate of 95% and a sensitivity rate of 61%.
P. Vigano et al. (Obstet. Gynecol. [2000] 95:115-118) report that the soluble form of intercellular adhesion molecule 1 is released by uterine endometrium and such release correlates with the extent (number of implants) of endometriosis in patients. The authors suggest that soluble intercellular adhesion molecule 1 might be of value in evaluating the spread potential of refluxed endometrium. However, soluble intercellular adhesion molecule 1 is also known to be released in other disease states so the potential value of this protein as a marker may be diminished somewhat.
J. Mahnke et al. (Fertil. Sterl. [2000] 73:166-170) evaluated VEGF and IL-6 levels in peritoneal fluid of women with endometriosis and found them to be elevated in patients with advanced disease. The levels of VEGF and IL-6 were lower in normal women and patients with milder disease. Nevertheless, the diagnostic value of these markers is suspect since at least VEGF is known to be a potent angiogenesis factor that is regulated by hypoxia in normal endometrium (A. M. Sharkey et al. J. Clin. Endocrinol. Metab. [2000] 85:402-409).
Matalliotakis and coworkers (Obstet. Gynecol. [2000] 95:810-813) found elevated levels of soluble CD23 in serum of women with endometriosis when compared to a control population. The CD23 levels decreased significantly during treatment with either danazol or leuprolide acetate. There seemed to be no correlation between soluble CD23 levels and the severity of endometriosis in the patients. As noted above for some of the other putative markers, CD23 has been associated with conditions linked to autoantibody production and levels of this protein are elevated in patients with autoimmune diseases.
Overall despite the substantial effort extended by numerous researchers, and also as reported in the publications reviewed above, no truly acceptable marker for endometriosis has been discovered. Yet, the physical and economic impact of the disease, and the difficulty in diagnosing the disorder dictate that the search for suitable markers be continued. Consequently, the activities disclosed in this invention were undertaken to identify markers of endometriosis that can aid physicians in monitoring patients with this illness. Other groups have performed such projects and these discoveries are the subject of numerous patent documents, which differ substantially from the discovery of the ME-5, ME-2 and EPP2 markers described in this invention. In U.S. Patent application 2003/0032044 there is a description of methods for generally detecting reproductive tract disorders by measuring the levels of interleukins IL-13 and IL-15 in specimens. Another U.S. Patent application 2002/0192647 proposes a process for diagnosing angiogenic diseases by measuring a single nucleotide polymorphism in the VEGFR-1 gene. Endometriosis is categorized as one of this group of angiogenic diseases, but it was not the subject of any of the claims. Patent applications 2001/046713 and 2001/044158 describe a method for diagnosis of endometriosis by detecting anti-Tomsen-Frienenreich antibodies in specimens. An issued U.S. Pat. No. 6,376,201 illustrates the use of major histocompatibility complex-class I antigens in diagnosing endometriosis and forming the basis of the Metrio Test as described above. In this patent the MHC-class I antigens are detected in specimens with specific monoclonal antibodies and similar disclosures were described in U.S. Pat. No. 5,618,680 and W.O. 0043789. A method for diagnosing endometriosis is described in U.S. Pat. No. 6,540,980 that involves measurement of eosinophil peroxidase levels. In U.S. Pat. No. 6,525,187 is described an apparently novel, marker of endometriosis which is the target of autoantibodies present in patient serum. Another method for diagnosis of endometriosis is disclosed in U.S. Pat. No. 6,387,629 and this is based upon the measurement of the protease cathepsin S in a clinical sample. A gene encoding an endometrial bleeding associated factor (ebaf) is described in U.S. Pat. No. 6,294,662 and this gene could be useful for diagnosis of endometriosis. However the ebaf gene seems to have better utility in the early diagnosis of selected carcinomas (colon, ovaries, or testis) in a human. In U.S. Pat. No. 5,877,284 another potential marker of endometriosis is described. This marker is a small soluble protein isolated by affinity chromatography from the peritoneal fluid of women with endometriosis, and the protein has chemotactic activity to neutrophils and macrophages. A process for monitoring human endometrial functions is described in U.S. Pat. No. 4,489,166 and it involves the quantitative measurement of progestagen-associated endometrial protein (PEP) in a clinical sample. European Patent No. 1191107 describes a method for diagnosis of endometriosis by measuring a reduction in the levels of one of a group of 15 different human genes. An immunoassay process is described in European Patent No. 0387027 which establishes endometriosis in a patient by evaluating a specimen with an anti-endometriosis monoclonal antibody. A method is described in W.O. 0063675 for diagnosis of endometriosis by measuring increased levels of endometriosis factor in biological fluids of a patient. W.O. 9963116 provides for a method of diagnosing endometriosis by measuring increases in the amount of prothymosin in endometriotic tissue.
U.S. Pat. No. 6,531,277 discloses an endometriosis-specific secretory protein. The document characterized and disclosed human ENDO-1 that is produced by stromal cells of endometriotic tissue. The ENDO-1 protein is 40 to 50 kilodaltons in molecular weight and has an isoelectric point of 4.0 to 5.5. The claims of the document are concerned primarily with a molecular diagnostic assay measuring differences in expression of ENDO-1 mRNA in endometriosis tissue samples. In a related application U.S. 2002/0009718 the invention is extended for measurement of the ENDO-1 glycoprotein in patient samples using immunoassay to establish the presence of endometriosis. Nevertheless, the characteristics of ENDO-1 presented in these documents suggest that it is considerably different from the markers described in the present invention. For example when measured by SDS PAGE and Western blotting the ME-5, ME-2, and EPP2 proteins are about 38, 49, and 9 kilodaltons in size, respectively. Only the ME-2 marker is within the range specified for ENDO-1, but ME-2 has an isoelectric point of 8.8 so it is not a related protein. Also, the isoelectric points of the ME-5 and EPP2 antigens are calculated at 5.7 and 12.5, respectively, which are also well above the range of values specified for the ENDO-1 protein. Moreover the ENDO-1 marker is a member of the haptogloblin family of proteins, but nucleic acid and amino acid sequence comparisons show that the ME-5, ME-2, and EPP2 markers are not related to this family of proteins.
In yet another separate disclosure, U.S. Pat. No. 5,843,673 specifies a method of screening for endometriosis in women by measuring a reduction in the amounts of a 28 to 32 kilodalton molecular weight glycoprotein in peritoneal fluid or serum samples. The protein possesses an isoelectric point of 7.0 to 9.0 and is secreted specifically by stromal cells of endometriotic origin. The glycoprotein disclosed in the document is related to tissue inhibitor of metaloproteinases-1 (TIMP-1) by virtue of amino acid sequence identity measured in the amino terminal region of protein. In the patent it is shown that endometriosis is indicated in a patient who has reduced levels of TIMP-1 present in serum or peritoneal fluid. The ME-5, ME-2, and EPP2 proteins of this invention are not related to TIMP-1 and they have no measurable protein or nucleic acid homology to this family of proteins. In addition, and as noted above, the biochemical properties of the ME-5, ME-2, and EPP2 proteins differ from those of TIMP-1 and each is considerably larger or smaller (at 38, 49, or 9 kilodaltons, respectively) than the range given for TIMP-1. While the isoelectric point of ME-2 is at the upper range of that of TIMP-1, the isoelectric point of ME-5 is 5.7 and EPP2 is 12.5 which are much different.
Another disclosure of protein agents implicated in endometriosis is contained in the document WO 01/32920 in which it is assumed that a total of 33 genes and their protein products are associated with the disease. These putative endometriosis markers were identified by comparing the pattern of gene expression in diseased endometrium relative to that of normal tissue. This differential display reverse transcriptase polymerase chain reaction employed in the document is a purely genetic screening approach designed to identify disease-associated genes based upon differences in the expression levels of mRNAs. The mRNA populations compared are usually normal healthy endometrium and the diseased counterpart, ideally both isolated from a single patient suffering from the illness. This technology ignores the functional activity of the proteins encoded by the mRNAs, and does not interrogate specimens based on disease hallmarks, symptoms, or the body's response to the illness. The latter strategies are arguably better approaches for marker discovery as discussed below. The individual nucleic acid sequences identified in the document fall into the general groups of; protease or protease inhibitor, tumor suppressor protein, immune system proteins, inflammatory response proteins, enzymes, lipid binding proteins, transcription factors, and matrix or cell adhesion molecules. All of the genes in WO 01/32920 are known and the nucleic acid sequences appear in the public databases allowing them to be identified. The individual nucleic acid sequences identified and implicated as somehow being involved in endometriosis are: cathepsin D, AEBP-1, stromelysin-3, cystatin B, protease inhibitor 1, sFRP4, gelsolin, IGFBP-3, dual specificity phosphatase 1, PAEP, immunoglobulin A chain, ferritin, complement component 3, pro-alpha-1 type III collagen, proline 4-hydroxylase, alpha-2 type I collagen, claudin-4, melanoma adhesion protein, procollagen C-endopeptidase enhancer, nascent-polypeptide-associated complex alpha polypeptide, elongation factor 1 alpha (EF-1α), vitamin D3 25 hydroxylase, CSRP-1, steroidogenic acute regulatory protein, apolipoprotein E, transcobalamin II, prosaposin, early growth response 1 (EGR1), ribosomal protein S6, adenosine deaminase RNA-specific protein, RAD21, guanine nucleotide binding protein beta polypeptide 2-like 1 (RACK1), and podocalyxin (and see references within WO 01/32920). Overall the diagnosis of endometriosis with the above agents would involve assessing the level of expression of the gene. The ME-5, ME-2, and EPP2 proteins and the nucleic acids described in this invention are also known and appear in the databases (see Example 1, below). However, none of the ME-5, ME-2, or EPP2 sequences fall into any of the groups listed above nor do they correspond to any of the designated agents either by computer-assisted homology comparison or predicted function based upon the presence of recognizable motifs present in the protein sequence. A similar gene expression-based strategy was employed in the discoveries documented by S. Baban et al. in US Patent Application 2002/0127555 in which 14 genes were found to be overexpressed in endometriosis patients relative to disease-free females. The overexpressed genes were NADH dehydrogenase, hUCCi, Paralemmin, citrate transport protein. HIF1-alpha, ARNT, Glut-1, MnSOD, GPx, ATP synthase, c-jun, Cx43, HSP 70, and cox2. In addition, 19 genes were reported in this document to be underexpressed in endometriosis patients relative to disease-free females. The genes underexpressed in diseased endometrial tissues were Cap43, RNA helicase, CO3, FKHR, AK3, catalase, GST, eNOS, 12S rRNA, T1227H, CO2, aconitase, ANT-1, Bcl-2, COUP-TF, IL-1 beta, HSP 90, GPx4, and GRP78. Yet another gene expression strategy was described by H. Hess-Stumpp et al. In US Patent Application 2003/0077589 resulting in the discovery of 15 genes that are overexpressed in endometriosis. The overexpressed genes were fibronectin, IGFBP-2, transmembrane receptor PTK7, platelet-derived growth factor alpha, collagen type XVIII alpha 1, subtilisin-like protein (PACE4), laminin M chain (merosin), elastin, collagen type IV alpha 2, p27 interferon alpha-inducible gene, reticulocalbin, aldehyde dehydrogenase 6, gravin, nidogen, and phospholipase C epsilon. Again, as stated above, the ME-5, ME-2, and EPP2 protein and nucleic acid sequences are not related to any of the genes described in the latter two patents.
Taken together and comparing the results of these three documents, it is interesting that all of them used similar but not identical gene expression strategies to identify a total of 62 genes which are overexpressed in endometriosis and 19 genes that are underexpressed. The implication is therefore that the 81 described genes are related to or involved in endometrial disease. Surprisingly, among these three independent studies, no single human gene or class of genes was consistently found to be associated with endometriosis. Ostensibly if a gene were overexpressed because of changes occurring in endometriosis tissue relative to the normal counterpart, then it would be expected to reproducibly be identified in all studies that assess the gene expression profile of diseased tissue. This does not seem to occur in the otherwise well-designed projects, and brings into question strategies for marker discovery based only on gene expression profiling technologies.
The document WO 94/28021 describes endometrial proteins, antigenic compounds, and methods of detecting endometriosis. The disclosure encompasses endometriosis-specific proteins defined by molecular weight and isoelectric point. Many of the claims presented are based only on size, but others specify a molecular weight and isoelectric point. The principal endometriosis antigen of the document and which is described in the initial claim has a molecular weight of 64 kilodaltons and an isoelectric point of 3.5. The antigen is used to measure antibodies in specimens obtained from endometriosis patients and also can itself be measured directly for its presence in patient samples. In addition, a larger molecular weight endometriosis protein of 94 kilodaltons with an isoelectric point of 3.5 is also described presumably to be used in the same formats as the smaller antigen. The document also claims nucleic acids for these proteins, however these sequences do not appear in enough detail to allow for comparison to the ME-5, ME-2, and EPP2 protein and nucleic acids of this invention. A small amount of amino acid sequence is presented in WO 94/28021, but there are only 17 residues shown in the document and of these over half are ambiguous. Although similar applications are envisioned for the ME-5, ME-2, and EPP2 protein described in this invention, the antigens described above do not compare in any reported properties to those of the three endometriosis antigens presented here. Initially, none of the unambiguous residues of amino terminal protein sequence are present in the corresponding regions of ME-5, ME-2, and EPP2. In addition, the ME-5, ME-2, and EPP2 proteins are 38, 49, and 9 kilodaltons in size, which are considerably smaller than the antigens described in the document outlined above. Moreover the isoelectric points of ME-5, ME-2, and EPP2 are 5.7, 8.8, and 12.5 which are considerably greater than described for the other proteins. It must be concluded that the endometrial ME-5, ME-2, and EPP2 antigens of this invention have little in common with the proteins described in WO 94/28021.
Methods and reagents for diagnosis of endometriosis are described in NZ 232801 (also application EP-A-0 387 027) essentially by measuring an endometriosis antigen in a patient specimen using an anti-endometriosis antibody. Various antigens are described in the document ranging in molecular weight from 50 to 173 kilodaltons but no additional characterization of the proteins was performed. These proteins were isolated as a mixture from the culture medium and cytoplasm of 2774 ovarian carcinoma cells, and can be obtained from other cultured cell lines as well. Also described in the disclosure is an anti-endometrial antibody, which is a human IgM monoclonal originally isolated because it reacted with ovarian cancer-associated antigens. Isolation of the antibody was apparently through a set of activities that were unrelated to endometriosis and the ovarian cancer antigen targets apparently were not well characterized. The antibody was made by fusion of patient lymphocytes with a heteromyeloma, and apparently the reactivity of the monoclonal with endometrial antigens was discovered subsequently. Regardless, based on the criteria presented it is unlikely that any of the proteins of NZ 232801 are the same as the smaller ME-5, ME-2, and EPP2 proteins of this invention.
Another series of endometrial antigens reactive with anti-endometrial antibodies is described in WO 92/18535 and these are also characterized by molecular weight on SDS PAGE analysis. The described protein antigen fragments were isolated from the cytoplasm of epithelial adenocarcinoma cells and are described as useful for detection of endometrial antibodies which are indicative of endometriosis. The antigens are cytoplasmic proteins with sizes of 63 to 67, 33 to 37, 40 to 44, 31 to 35, and 57 to 64 kilodaltons. The designations likely refer to a single protein species, but the size ranges were presented in the document to reflect the inherent inaccuracy (±10%) for the SDS PAGE assay method used. Apparently the preferred proteins for use are the 33 to 37, 40 to 44, and the 57 to 59 kilodalton proteins. The 33 to 37 and 40 to 44 proteins seemed to be present in most of the cell lines that were studied in the document for use as sources of antigen, while the 57 to 59 protein fragments originates from the T47D breast carcinoma cell line. The document describes the use of these proteins individually (or mixed) immobilized on solid support to measure endometrial antibodies. Of course similar applications are envisioned for the ME-5, ME-2, and EPP2 antigens, however with the exception of possibly the 33 to 37 kilodalton fragments there is little else presented in this document that compares to disclosures in WO 92/18535.